Interactive plots: topic introduction

In this part of the course, we will cover the following concepts:

• Create correlation plots, column plots, box plots
• Save and view interactive plots by using htmlwidgets library

Warm-up trivia

• Before moving forward, let’s review the series used in highcharter

• Below are 2 columns:

  • Column A shows the highcharter series type
  • Column B is the plot type

• Pair the correct series type and plot type, and share your answers in chat

Review trivia answers

• column is for creating a bar plot (1-b)
• area is for a density plot (2-a)
• bar is for creating a horizontal bar plot (3-c)

Module completion checklist

Creating plots with highcharter

• We will work with the healthcare stroke dataset to create interactive visualizations
• We will start with a boxplot to visualize the distribution of bmi based on the smoking status
• We will then make a multiple-column plot to help compare different summary statistics by variable
• Finally, we will create a correlation plot to assess the strength of the relationships between variables

Directory settings

• In order to maximize the efficiency of your workflow, use the box package and encode your directory structure into variables

• Let the main_dir be the variable corresponding to your materials folder

# Set `main_dir` to the location of your materials folder.

path = box::file()
main_dir = dirname(dirname(path))

Directory settings (cont’d)

# Make `data_dir` from the `main_dir` and
# remainder of the path to data directory.
data_dir = paste0(main_dir, "/data")
# Make `plots_dir` from the `main_dir` and
# remainder of the path to plots directory.
plot_dir = paste0(main_dir, "/plots")

Introducing HDS data set

• We will explore a dataset called healthcare-dataset-stroke-data

• This dataset contains information about age, gender, hypertension, bmi, and other parameters to know the chances of getting a stroke

• The goal is to understand how different variables in the dataset affect the chances of a person suffering from a stroke

• The dataset has 12 characteristics (columns), of which:

  • 10 columns relate to the quality and characteristics of the life of different people
  • The stroke column represents whether the people had a stroke or not

Load HDS dataset

# Read CSV file called "healthcare-dataset-stroke-data.csv"
HDS = read.csv(file = file.path(data_dir,"/healthcare-dataset-stroke-data.csv"), #<- provide file path
               header = TRUE,            #<- if file has header set to TRUE
               stringsAsFactors = FALSE) #<- read strings as characters, not as factors

Variables in the data

Prepare data

• But before sub-setting the data, let’s handle the missing data in the dataset
• Then convert bmi into a numeric column followed by imputing the missing values with the mean

HDS$bmi <- as.numeric(as.character(HDS$bmi)) ##converting bmi column to numeric
# NA imputation
# we can use is.na() function to know about NA values
HDS$bmi[is.na(HDS$bmi)]<-mean(HDS$bmi,na.rm=TRUE) # Replacing na values of bmi column with it's mean bmi

Subset data

• Now, we will subset our data to include the columns we will use to build the box plot

# Let's make a vector of column indices we would like to save.
column_ids= select(HDS, age, bmi, smoking_status,work_type)
# Let's save the subset into a new variable.
HDS_subset = column_ids

Highcharter boxplot: hcboxplot()

library(highcharter)
library(tidyverse)

?hcboxplot

hcboxplot(x = Numeric_data_vector,
          var = Categorical_data_vector,
          ...)

Highcharter boxplot: hcboxplot (cont’d)

# Construct an interactive boxplot.
boxplot_interactive =  
  hcboxplot(x = HDS_subset$bmi,
            var = HDS_subset$work_type) %>%
  hc_title(text = "HDS data variables")

boxplot_interactive

Highcharter Plotting Options

Customize hcboxplot with hc_plotOptions

# Enhance original boxplot with color options.
boxplot_interactive = boxplot_interactive %>% 
  hc_plotOptions(   #<- plot options
    boxplot = list(     #<- for boxplot 
      colorByPoint = TRUE)) #<- color each box

boxplot_interactive

Column plot: prepare data

• Let’s create a new subset of our data with the columns needed to create a column plot

column_ids= select(HDS, age,bmi,avg_glucose_level)
# Let's save the subset into a new variable.
HDS_subset_col = column_ids

Column plot: prepare data (contd)

# Create data summary.
HDS_summary = summary(HDS_subset_col)

# Save it as a data frame.
HDS_summary = as.data.frame(HDS_summary)

# Inspect the data.
head(HDS_summary)

  Var1     Var2            Freq
1           age Min.   : 0.08  
2           age 1st Qu.:25.00  
3           age Median :45.00  
4           age Mean   :43.23  
5           age 3rd Qu.:61.00  
6           age Max.   :82.00  

# Remove an empty variable.
HDS_summary$Var1 = NULL

# Rename remaining columns.
colnames(HDS_summary) = c("Variable", 
                          "Summary")
# Inspect updated data.
head(HDS_summary)

  Variable         Summary
1      age Min.   : 0.08  
2      age 1st Qu.:25.00  
3      age Median :45.00  
4      age Mean   :43.23  
5      age 3rd Qu.:61.00  
6      age Max.   :82.00  

Column plot: prepare data (contd)

• Let’s separate the values from the statistics using tidyr’s separate() function

# Separate `Summary` column into 2 columns.
HDS_summary = HDS_summary %>%              #<- set original data 
  separate(Summary,                        #<- separate `Summary` variable 
           into = c("Statistic", "Value"), #<- into 2 columns: `Statistic`, `Value`
           sep = ":",                      #<- set separating character
           convert = TRUE)                 #<- where applicable convert data (to numeric)


# Inspect the first few entries in data.
head(HDS_summary)

  Variable Statistic Value
1      age   Min.     0.08
2      age   1st Qu. 25.00
3      age   Median  45.00
4      age   Mean    43.23
5      age   3rd Qu. 61.00
6      age   Max.    82.00

# Inspect total number of rows in data including NAs.
nrow(HDS_summary)

[1] 18

Column plot: create plot

• Then, clean the data by removing NA observations and build the chart

# Inspect `Value` column for `NAs`.
which(is.na(HDS_summary$Value) == TRUE)

integer(0)

HDS_summary = subset(HDS_summary, !is.na(Value))

nrow(HDS_summary)

[1] 18

# Construct the summary chart.
HDS_summary_interactive = 
  hchart(HDS_summary,             #<- set data
         "column",                #<- set type (`column` in highcharts)
         hcaes(x = Statistic,     #<- arrange `Statistics` across x-axis
               y = Value,         #<- map `Value` of each `Statistic` to y-axis
               group = Variable)) #<- group columns by `Variable`

Column plot: display plot

HDS_summary_interactive

Column plot: customize tooltip

• While comparing each variable’s summary statistics, it would be convenient for the tooltip to contain information about the group of variables i.e age, bmi, and avg_glucose_level rather than than the individual variables
• We can control different tooltip options of the chart using the hc_tooltip() option
• The shared option is often used to share a tooltip between members of a group

# Adjust tooltip options by piping `hc_tooltip` to base plot.
HDS_summary_interactive = HDS_summary_interactive %>% 
  hc_tooltip(shared = TRUE)  %>%               #<- `shared` needs to be set to `TRUE`
  hc_title(text = "HDS data variable summary") #<- add title to your plot

Column plot: customize tooltip (cont’d)

HDS_summary_interactive

Module completion checklist

hchart: correlation plot

• Certain charts require less data processing within the highcharter package
• For instance, if we pass hchart() a correlation matrix, it will recognize the data type and create a correlation plot in response
• No other arguments are necessary to create a basic plot

# Compute a correlation matrix for the first 
# 4 variables in our data.
cor_matrix = cor(HDS_subset_col[, 1:3])

# Construct a correlation plot by 
# simply giving the plotting function
# a correlation matrix.
correlation_interactive = hchart(cor_matrix) %>%
  # Add title to the plot.
  hc_title(text = "HDS data: correlation")

hchart: correlation plot (cont’d)

correlation_interactive

Save interactive plots: htmlwidgets

# Install `htmlwidgets` package.
install.packages("htmlwidgets")

# Load the library.
library(htmlwidgets)

# View documentation.
library(help = "htmlwidgets")

Save interactive plots: htmlwidgets (cont’d)

• We can save widgets to a dedicated plot directory

# Set working directory to where you save plots.
setwd(plot_dir)

# Save desired interactive plot to an HTML file.
saveWidget(scatter_interactive,        #<- plot object to save
           "interactive_scatterplot.html", #<- name of file to where the plot is to be saved
           selfcontained = TRUE)              #<- set `selfcontained` to TRUE, so that 
                                              #   all necessary files and scripts are embedded 
                                              #   into the HTML file itself  

Knowledge check

Link: Click here to complete the knowledge check

Exercise

Module completion checklist

Interactive plots: topic summary

In this part of the course, we have covered:

• Creating correlation plots, column plots, box plots
• Saving and viewing interactive plots by using htmlwidgets library

Congratulations on completing this module!